An Empirical Research On The Collegiate Learning Network Structure On The Basis Of Social Network

With the development of digital revolution, the collegiate learning behavior has transformed dramatically. The collegiate learning behavior has transmitted from the traditional "individual learning" to "group learning" and the learning structure has transmitted from the "door-closed-mode" to the "door-opened-mode". The connection of learning becomes a key issue to impact the effectiveness of learning. Thus the thesis defines the "learning network" as a social network that is constructed during the learning process, aiming at gaining knowledge for the purpose of information delivering.At present, research on the social network structure of collegiate learning network is in its infant stage. Although some researchers have investigated either the sociality of collegiate learning or the whole learning network structure, there lacks an academic coverage on the theme of whole collegiate learning network. Based on the Connectivism Learning Theory, the thesis embeds the collegiate learning behavior into social network, and from the perspective of structuralism, the thesis utilizes the analytical paradigm of social network and other relevant methodologies to investigate the whole structure of collegiate learning network.The thesis consists of three main parts:Firstly, the thesis constructs a theoretical framework to analyze the corresponding network structure of collegiate learning behavior. Drawn on the analytical paradigm of social network, the thesis originates from the objective existence of collegiate learning network, abstractively identifies the essential features of learning network, proposes a set of workable methods of evaluation and analysis and then constructs an analytical model of learning network structure.Secondly, the thesis investigates the structural features of collegiate learning network with both qualitative and quantitative methods. The network structure directly influences the behavior of person of action, therefore, utilizing the analytical method of social network, the thesis investigates the features of learning network structure from and from three different perspective, both macro, medium and micro levels. Drawing from the theoretical analysis and relying on the sampling data, the thesis undertakes an exploratory research and evaluates those specific structural variables.Thirdly, the thesis constructs a model of the whole collegiate learning network structure. The usual approach in research on management is to identify both the independent and dependent variables of research subjects, and tries to build relationships between them. The purpose of constructing the model is to discover the very parameters which have significant correlation with the relating network, thus to deduce whether a substructure is more likely to exist in the learning network than accidental appears.The thesis has gained the following research findings:Firstly, the thesis constructs an analytical model of collegiate learning network structure, accordingly the thesis undertakes a qualitative analysis on the collegiate learning network and finalizes the visual expression of the whole collegiate learning network. An empirical research verifies that the proposed structural analytical model can ideally be utilized in analyzing the collegiate learning network structure.Secondly, combining both the structural analytical model and qualitative analysis, the thesis evaluates both the structural features of the practical collegiate learning network and its variables, thus it identifies the features of the learning network.Utilizing the exponential random graph model, the thesis constructs a model of collegiate learning network structure and identifies seven structural variables that have significant effect on the whole network structure. Among them, three structural variables namely reciprocity, transitivity and block density can preferably interpret the whole learning interbehavior.In terms of the collegiate learning network structure, the thesis draws the main conclusions as follows,Firstly, the relational scheme of collegiate learning is a complex network with multi-paths and multi-hierarchies.Secondly, each student averagely will build learning relationships with four nodes, but the numbers of nodes differentiate much from each other.Thirdly, in the collegiate learning network, the transitivity is strong, the transmission channel is open, the information exchange is fast and the density of a network is high, but the correlation is relatively low.Fourthly, the network contains several cohesive subgroups. The subgroups have strong internal connections but the connection to other subgroups is loose.Fifthly, there are many Trio-learning groups in the network. Among them, the numbers of instable star schema are in the majority, while the numbers of stable annular mode and the reticulation pattern are the least.Sixthly, the selection of external nodes appears to be of diversity. "Friends", "search engine", "none classmates" and "teachers" are the four most selected nodes.In terms of the innovation and theoretical contributions, the thesis concludes from the following three perspectives,Firstly, the thesis proposes a new research thought. The thesis embeds the collegiate learning behavior into social network, which regards the collegiate learning relationships as a group to undertake a holistic research. This breaks the methodological limitations of the former research merely on individuals, and points out a new research thought for investigating the collegiate learning behavior.Secondly, the thesis presents a new opinion on learning. The thesis firmly believes that the relationships in the collegiate learning process not only are impacted by its personal attribute, but also are influenced by its learning environmental status, that is the learning network structure. The collegiate learning behavior relates to its position in the learning network and meanwhile the learning network both restricts the collegiate learning behavior and interprets its behavioral features.Thirdly, the thesis constructs two new interpretive models. In order to investigate the collegiate learning network structure, borrowed from other former researches, the thesis constructs an analytical model of the contemporary collegiate learning network structure, and the research findings have proved the effectiveness of the analytical model. Meanwhile, based on both qualitative and quantitative analysis, the thesis builds a model of collegiate network structure utilizing the exponential random graph model, which perfectly explains the causal relationship between the variables of network structure and the probability of occurrence of collegiate learning relationships.The issue of collegiate learning relationships is both an enormous and complex research field worthy of further investigation. The thesis has undertaken a preliminary discuss and the research design, the methodology and sampling strategy have some limitations. For example, the categorization of collegiate learning relationship needs to be expanded and improved, and the causal model between learning network structure and learning effect is not investigated yet. The improvement and complement to the existing limitations will guild the further research in the future.